Halamka: Decision support, care management key to accelerating precision medicine
Beth Israel Deaconess Medical Center CIO John Halamka, MD, has written and presented often over the past few years about his wife Kathy's cancer treatment and the ways precision medicine techniques have helped guide her care plan.
In a recent HIMSS Learning Center webinar, he shared some lessons learned from a recent chapter in that journey and offered some optimistic words about the timeline for improving the delivery and reimbursement of personalized treatments.
In Halamka's telling, effective precision medicine entails and requires much more than just the genomic science most commonly associated with the term. It depends on patient and family engagement, social determinant factors, easily interoperable clinical decision support tools and more.
For one example, he cited a letter Kathy Halamka received letter from her insurer, the highly rated Harvard Pilgrim, explaining that it was suddenly denying coverage for the current dosage one of her ongoing estrogen-suppressing drugs because it had found an old research paper that showed a dosage of half that amount would be equally as effective.
Halamka, who will be speaking Monday at the HIMSS Big Data and Healthcare Analytics Forum in Boston, wrote extensively about the episode on his blog this past month.
When asked, Harvard Pilgrim conceded that it hadn't consulted Kathy's records, protocols or preferences to determine whether the choice was right or justified, and hadn't reviewed anything about Kathy's personal characteristics that "would suggest that potentially 22.5 mg was the right dose," instead of the 11.25 mg it was willing to reimburse.
Partly perhaps owing to his position as one of the most recognizable figures in healthcare, Halamka was able to consult with Harvard Pilgrim at a high level and asked it to consider embracing a more evidence-based approach.
"Payers and providers need to relate to each other using an evidence base to connect the two so we're delivering the right care for quality and outcomes," he said. "In many parts of the country payers and providers don't like each other very much. And the notion of exchanging clinical data from provider to payer for care management and precision medicine is still a psychology problem, not so much a technology problem."
Thankfully, Harvard Pilgrim is one of the most forward-thinking health plans in the U.S. It agreed, after further discussion, to stop sending letters like the one Kathy received, said Halamka.
Instead, future changes in approved treatment plans would be explained by saying, "We have, through evidence, determined that you might be on a medicine that's not optimal for you," he said. "It might cause too many side-effects, might not be as effective as it could be."
There would be phone calls, communications with care teams, reviews of patient records, their protocols and perhaps their genetic makeups. If it's decided on both sides of the provider-payer equation that there might be a better dosage, the patient might eventually get an e-prescription from the care team that explaining why it's being changed.
A process like that incorporates the essential elements of precision medicine, said Halamka: "Who is the patient, what are their preferences, what is their genetic makeup, what is the evidence that suggests that one treatment is better than another for that individual?"
We're getting there, but "it's still early" and it may be some time before "we exactly get it right," he said. But perhaps not as much time as some think.
Imagine if ...
"Here's how it should have worked," said Halamka. "Imagine if, in the EHR, because clinicians hate going to some website or separate application, when the clinician goes to order the information, it's sent to a cloud-hosted decision support provider," along with some minimal amount of patient information that might be relevant: age and gender, diagnosis, some lab values.
"Then the decision support service provider returns some pleasing answer," he said. "'For this particular patient this particular dosage will be maximally effective with the fewest side-effects. Instead of having to do this in a post-coordinated fashion, six months after the patient is already on a protocol, you're doing it as the order is initiated, which gives the opportunity for the patient, doctor and decision support to all be in the same room at the same time.
If that sounds too good to be true, it's much closer to being common practice that many might assume.
"Lest you think that is something five years hence," said Halamka, "as we're starting to see, more and more of our EHR vendors – Cerner, Epic, Meditech, athenahealth, eClinicalWorks – are embracing the idea of using the FHIR standard to connect to third-party applications and knowledge sources."
He pointed to the FHIR-based CDS Hooks standard, which is "exactly designed for the EHR to contact some external knowledge provider and consume the result inside the EHR workflow."
The clinicians get guidance showing possible treatment choices and objective rankings of safety, quality, efficiency, cost, and availability. And then the clinicians and their patients have a discussion and, via shared decision-making, develop a care plan."
And array of open-source apps are increasingly showing their worth when it comes to communicating care plans to the patient and relaying patient-generated healthcare data back to the provider, of course, and that is where the energy will be on this front, said Halamka.
"EHR vendors are doing great – they're doing everything we need for regulatory compliance and revenue cycle," he said. "But they are not likely to be the place where radical new innovation occurs, in AI and matching learning, or the ability to provide this very evidence-based skinny downset of actionable items for clinicians at the point of care," he said.
Halamka predicts that we'll be seeing much more commonplace "production-level FHIR exchanges" between EHRs and cloud-hosted clinical decision support providers "in the next 12 months."
Hurdles ahead, but big innovations are coming
In the meantime, he sees other innovations sprouting that will enable new momentum toward more personalized treatment plans.
For instance, he predicted that consensus will finally soon start to emerge on nationwide patient-matching strategy. That will be essential "if we're going to deliver precision medicine as I've described – data about you in a more continuous rather than episodic fashion, and knitting it together," said Halamka.
"I do think that in the next year, you'll see not a national identifier, but a strategy, guidelines" – it could be biometrics, it could be any number of different approaches being workshopped by an array of groups such as CHIME, Pew Charitable Trusts, the CommonWell Health Alliance – that will enable patients to be easily identified across care settings.
Accountable care, he said, is helping to force that "urgency to change," said Halamka, and "the drive to value-based-purchasing will help accelerate the drive to precision medicine."
Add to that a healthcare landscape that figures to be (relatively) unburdened by federal regulations after a decade dealing with HITECH, ICD-10, the HIPAA Omnibus, ACA, etc.
"I think we can say the next three years will probably have fewer legislative and regulatory efforts," said Halamka. "That means the private sector are going to guide where we're going forward. And it gives us the breathing room to drive the precision medicine innovation I've been talking about."
There are challenges ahead, of course: The tools patients and families use to find their way around the care delivery system – portals, primarily, are suboptimal, to say the least. And on the provider side, clinicians are hardly much more enamored of their EHRs. The infrastructure for information exchange also, clearly, leaves lots to be desired.
But Halamka says he's optimistic about the prospects of better integrating care plans and clinical pathways into the tools of the trade (EHRs, PHRs and apps) "so the patient, the family and the doctor all have a single playbook to work from."
That's the name of the game, he said. It's going to "reduce errors and cost, include appropriate peer-rated evidence, result in clear action for the care team."
And in a value-based work, "it's a team-based sport," said Halamka. "It isn't just a PCP and a patient. It's going to be the family, the nurse, the pharmacist, the social worker. All folks working together it's really important you have a common care plan for disease state and patient to work against."
Progress is being made at Beth Israel Deaconess Medical Center, he said, which is developing technologies and strategies to enable more personalized interventions – such as a pilot to deliver to patients' phone, in real time, actionable items for their specific disease state and conditions, and then allow them to report back on their progress using an array of IoT devices.
BIDMC is an innovation outlier – it's the last hospital in America to self-build its core electronic health record – but "I can assure you that Epic and Cerner are busy at work creating very similar functionality," he said.
"This is the kind of integration with precision medicine care planning, workflow with doctors patients and families that we all want," said Halamka.
"I believe that the promise of precision medicine is real," he said. "And our experience with patient family engagement, the emerging promise of machine learning and IoT connectivity, combined with evidence, will get us there – not in five years, but in one."